Two Buffalo philosophers will argue on November 3 that AI will not take your job
Barry Smith and John Beverley will give a free public talk at the University at Buffalo; Smith calls AI's promise wildly exaggerated.
In brief
A free talk says AI is overhyped
Two Buffalo philosophers speak on November 3 at 5 p.m. They argue AI will not take jobs. The claims come from the university's own release, not outside testing.
The University at Buffalo announced a free public talk called "Why AI Will Not Steal Your Job." The announcement is a university news release dated October 27, 2026, so the claims below are the speakers' own, not tested findings.
When and where
Smith says AI does not typically excel at activities that would replace human beings.
The talk is at 5 p.m. on November 3 in Knox 104. Anyone can attend. UB's own UBNow page lists the same event.
Who is speaking
Barry Smith is a SUNY Distinguished Professor of philosophy. John Beverley is an assistant professor of philosophy. He also directs UB's graduate program in applied ontology and is the deputy head of UB's institute for artificial intelligence and data science.
What they say
Smith's message to worried workers is short: don't be. He says AI does many things well. But much of human intelligence involves things that cannot be copied from the data used to train AI.
He also says AI does not usually excel at the activities that would replace people or take over where people stop.
The release says the talk will cover the real state of the AI industry, its effect on employment, and the kinds of jobs it is creating. It will also cover how students in STEM and non-STEM fields can prepare for careers.
The argument behind it
Smith is co-author of the book "Why Machines Will Not Rule the World." With co-author Jobst Landgrebe, he revisited the argument two years later in the article "Why Machines Will Still Not Rule the World." Their thesis is that no machine will ever reach artificial general intelligence, meaning AI that matches or beats humans at everything.
The release also names OpenAI and Anthropic as leaders in scaling. That is the bet that feeding ever more data into computers will eventually produce human-level machines. Smith says scaling has two flaws. It is very expensive, and the data comes from people, so it carries human faults.
The release also gives Smith's view of AI's economics: he says the losses are huge and growing because there is no market to buy the results, and that companies pay "$10 for $1 worth of output." Technology.org, which credits the university, says he also calls the bubble likely to pop.
What the sources do not show
The coverage comes from the university's release and reprints of it. No outside newsroom reporting was found. The sources give no data on jobs, and no one outside UB responds to the speakers' claims.